Ammonia (NH₃) emitted from animal feeding operations is a major air and water pollutant contributing to surface water eutrophication, soil acidity, and fine particulate matter (PM_2.5) formation [1,2]. Current estimates for livestock contribution to anthropogenic NH₃ in the U.S. are at 50% [1]. Some reports have indicated, however, that a significant portion of manure N lost during storage may be as non-NH₃ gases, such as dinitrogen gas (N₂) [3]. The latter authors suggested, for example, that N₂ emissions from swine lagoons are many times greater than emissions of NH₃. Emissions of N2 from cattle manure may be also high, particularly during the initial stage of manure storage when the bulk of urinary N is volatilized. If this is the case, mass balance, or other indirect approaches (i.e., not measuring NH₃ emissions directly; isotope, manure minerals:N ratios [4]) for estimating NH3 emissions may not be accounting for gaseous non-NH3-N losses and thus, NH3 emissions from cattle operations may be overestimated. For example, 25 and almost 50% of the daily N flow in dairy and beef cattle operations, respectively, were unaccounted as milk, daily body weight gain, or manure [2]. How much of this loss is NH₃ and how much non-NH₃-N is unknown. It is important to point out that N₂ is an inert gas and, unlike NH₃, is not considered an air pollutant.

Of the two major N pools in cattle (or most farm animals) manure, feces and urine, the latter (specifically, urinary urea in cattle) is generally considered to be the major source of emitted NH₃ [5]. Although the biological and biochemical ground for such an assumption is solid, there is surprisingly little experimental data to support it. For example, the conclusions of Bussink and Oenema [5] are primarily based on a study with soil application of synthetic urinary N compounds [6]. To our knowledge, only one study directly investigated urinary vs. fecal N contribution to volatile N emissions from animal manure [7]. Nitrogenous gas emissions from manure are to a large extent dependent on manure composition [2], which in turn depends on the animals' diet. Thus, it is important to quantify the actual contribution of urinary N to these emissions, particularly in the initial stages of manure storage when emissions are most intensive, which would allow for successful mitigation of manure emissions through dietary means.

A substantial part of mitigating manure emissions, including NH₃, is the availability of accurate and practical methods for estimating emissions. Direct measurement techniques are “the gold standard”, but are affected by a multitude of environmental factors (temperature, wind velocity; see later discussion) and are of limited value when, for example, the effect of diet on manure emissions is evaluated [2,8]. The U.S. Environmental Protection Agency (EPA) recently released data from the National Air Emissions Monitoring Study [9], in which gaseous emissions, including NH₃, from several commercial dairy operations were monitored. In this study, barn NH3 emissions varied from 4.6 (a California dairy) to 78 g/cow/day (a Washington dairy). Similar large variability in directly monitored NH₃ emissions from dairy farms (0.82 to 250 g NH₃/cow/day) or beef feedlots (50 to 283 g NH₃/animal/day) was reported in a recent literature review [2]. With such large variability, determining the specific effect of diet is practically impossible. Therefore, we have investigated indirect methods for estimating manure NH₃ emissions, utilizing minerals:N ratios and natural N isotope fractionation [4]. The isotope method appeared promising, however, the relationship between δ¹⁵N of manure and NH₃ volatilization is a dynamic process and longer monitoring periods are necessary to determine the usefulness of this approach for practical applications.

In this study, a series of laboratory experiments were conducted with the following objectives: (1) establish the relationship between manure N isotope composition and NH3 emissions beyond 10 days of storage; (2) investigate the potential contribution of nitrogenous gases other than NH₃ to manure N volatilization losses; and (3) determine the relative contribution of urinary- vs. fecal-N to NH₃ emissions during the initial stage of manure storage. We hypothesized that: (1) δ¹⁵N of NH₃ and manure N will continue to increase beyond 10 days and will reach a plateau; (2) non-NH₃ gases, such as N₂, may account for a significant portion of manure N losses, particularly during the initial storage phase; and (3) urinary urea-N is the primary source of NH₃-N emitted from cattle manure during the initial, most intensive, phase of manure N volatilization losses.

Dry matter and concentration of total and urea-N in manure used in this study varied significantly among experiments (Table 1). Manure N and specifically urea-N are important factors determining NH₃-N volatilization losses from cattle manure [2]. Manure in Exp. 2 had about 50 to 60% lower (P < 0.001) urea- and total-N concentrations compared with manure used in Exp. 1, 3, and 4. This led to significantly lower daily manure N losses in Exp. 2, compared with Exp. 1, 3, and 4. The highest (P < 0.001) daily N losses were in Exp. 4, which can be explained by the shorter duration of this experiment (10 days), compared with the other experiments (20 or 30 days). The most rapid loss of manure N and most intensive NH₃-N emissions occurred during the first 5 to 6 days (Figure 1A,C). This was matched by an equivalent rapid increase in NH3-N concentration in manure, reaching a peak at day 2 to 5. Initial concentration of ammonium in manure was negligible, but rapidly increased (to about 3 to 5 mg/mL manure) through day 5 in both Exp. 1 and 2 due to hydrolysis of urinary urea (data not shown). The much more rapid decline in manure urea-N concentration (Figure 1B) suggests that although urea hydrolysis took place immediately following mixing of feces and urine, NH₃-N volatilization was a slower process. As shown in Table 1 and Figure 1 (Panels A and C), the quantity and intensity of manure N losses and NH₃-N emissions were much lower in Exp. 2. As a proportion of manure N at day 0, N losses were similar (P > 0.05) between Exp. 1, 2, and 4, even though the duration of Exp. 2 was 30 days (compared with 20 or 10 days for Exp. 1 and 4, respectively). This again, emphasizes the importance of urinary urea-N concentration in the early stages of storage for the magnitude of NH₃-N losses from manure. The daily NH₃-N losses were the lowest (P < 0.001) in Exp. 2, but the highest as a proportion of manure N losses compared with the other experiments (Table 1). The lowest total recovery of manure N lost during the incubation was for Exp. 4, which had the shortest incubation length (10 days).

The higher manure N recovery as NH₃-N in Exp. 3 vs. Exp. 1 (both 20 days in length) can be related to the lower N concentration of manure in Exp. 3. The relationship between manure N concentration and manure N recovery as NH₃-N, which was linear and negative for Exp. 1, 2, and 3 [147.7 − 11.0 × N concentration in manure (%); R² = 0.86; P < 0.001], presents an interesting phenomenon. Recovery of manure N as NH₃-N captured in the acid trap was generally low. Recovery was even lower in the initial stages of the incubation (on average, 19.2 ± 0.72% during the first 3 days of incubation) suggesting that: (1) the acid trap did not effectively capture NH₃-N emitted from manure, particularly when emissions were most intensive, or (2) N was being lost from manure in forms other than NH₃-N. We have investigated the factors affecting the NH₃-N trapping efficiency of acid solutions and reported that efficiency decreases with increasing the amount of NH₃-N being emitted [13]. Decreased trapping efficiency, however, could not explain the large discrepancy between manure N losses and NH₃-N captured in the acid trap in Exp. 1 through 3. To further eliminate the NH₃-N trapping capacity of the acid solution as a factor for the low recovery of manure N lost during the incubation process, we conducted a series of experiments comparing the acid trap system with direct measurement of NH₃-N emitted from the incubation vessels using a photoacoustic gas analyzer INNOVA 1412 (AirTech Instruments, Ballerup, Denmark), which allowed continuous monitoring of NH₃-N concentration in the gas flowing out of the system. The conclusion from these experiments was that the 2 measurement methods gave similar NH₃-N recovery. For example, cumulative NH₃-N emissions were 108 vs. 121 mg in 24 h (SEM = 4.99; P = 0.10) and 128 vs. 136 mg in 72 h (SEM = 6.40; P = 0.50) for the acid trap and the INNOVA gas analyzer, respectively.

The possibility of a significant gaseous N loss, other than NH3-N, was further investigated. Nitrous oxide emission is expected to be negligible in conditions as those utilized in the current study due to the lack of nitrifying and denitrifying microorganisms in cattle feces [14] and relatively short storage time. Adviento-Borbe et al. [15] and Arriaga et al. [16], for example, reported insignificant N₂O emissions off the barn floor in dairy farms. In an experiment related to this study, N₂O emissions were negligible from dairy manure stored in laboratory conditions or during the first 100 h following soil application [17]. If sufficient time (at least 3 weeks) is allowed, however, cattle manure will generate N₂O in simulated storage conditions [18]. Bussink and Oenema [5] and Harper et al. [19] indicated that some N may be lost from lagoons/retention ponds via reduction of nitrate to N₂O and dinitrogen gas (N₂). A number of possible chemical and biological mechanisms may exist for formation of N₂ during manure storage [20] and such processes, including chemical, non-biological conversion of ammonium to N2 (termed “chemo-denitrification”) have been reported to be responsible for a significant amount of gaseous N losses from swine lagoons [3]. In a separate series of experiments we used pure argon gas (99.99%; GTS-Welco, Allentown, PA) instead of air to provide airflow in the manure storage system used in Exp. 1 through 4 and analyzed the composition of the gas flowing out of the system. Preliminary results from these experiments (data not shown) indicated very intensive N₂ emissions in the first 5 h of simulated manure storage, suggesting that N2 gas may represent a significant N loss in the initial stages of the manure storage process and likely accounts for a significant part of the N losses observed in the current study.

As discussed earlier, farm NH₃ emissions are influenced by important environmental factors and such data are not suitable for evaluating the impact of dietary mitigating strategies. For example, in a current on-farm project with 12 commercial Pennsylvania dairy farms, we monitored barn floor NH₃ emissions in spring and fall of Year 1 and then again in Year 2 of the project, after dietary crude protein concentrations were reduced by about 1%-unit [21]. On average, barn floor NH₃ emissions for the farms, in which the dietary protein reduction was documented by regular sampling, were reduced by about 65% (445 vs. 156, mg/m²/h). However, average air temperatures during the emission measurements were 14 °C and 5 °C, respectively. Thus, in this particular project, it was impossible to distinguish the effect of diet from the effect of environment. Manure samples from the same farms (collected, stored, processed, and analyzed as in the current study) showed unequivocally a reduction in laboratory NH₃-N emissions by about 36% for the low-protein period compared with the control, high-protein feeding period. Laboratory methods, naturally, have the limitation of not accounting for the environmental factors affecting emissions, but are useful in quantifying the effect of dietary manipulations on the gas-emitting potential of manure [2].

One of the objectives of the current study was to further investigate the relationship between NH₃-N volatilization and N isotope composition of manure. We first reported a significant N isotope fractionation in cattle manure during storage due to NH₃-N volatilization [4,22]. The isotope fractionation factor associated with NH₃-N volatilization is one of the highest in the N cycle (∼1.029, [23]), which would result, when conditions are favorable, in a rapid increase in δ¹⁵N of manure during storage. The process has been discussed in length [4]. Experiments 1 and 2 were used to determine N isotope ratios beyond the short incubation time utilized in our original studies [4,22].

Delta ¹⁵N of manure rapidly increased from 0.09 ± 0.36 (day 0) to 10.1 ± 0.42 (day 5) in Exp. 1 and from −1.12 ± 0.61 (day 0) to 5.99 ± 0.40 (day 5) in Exp. 2 (Figure 2A). This rapid increase in δ¹⁵N was due to the loss of depleted in ¹⁵N NH₃-N. Delta ¹⁵N of volatilized NH₃ was −22.5 ± 0.68 on day 1 in Exp. 1 and −15.1 ± 0.17 on day 2 (day 1 measurement was lost) in Exp. 2 and increased to −16.5 ± 0.09 (day 5) and −1.3 ± 2.55‰ (day 20) in Exp. 1 and to −14.9 ± 0.90 (day 5) and 2.38 ± 1.45‰ (day 30) in Exp. 2 (Figure 2B). As hydrolysis of urea to ammonium (which as Figure 1B shows is a very rapid process), NH₃-N volatilization, and N isotope fractionation take place, δ¹⁵N of NH₃-N and the dissolved in manure ammonium will continue to increase until the ammonium is exhausted and the NH₃-N obtains the δ¹⁵N value of the original ammonium. Indeed, as Figure 2B shows, δ¹⁵N of volatilized NH₃-N continued to increase in a sigmoid fashion through day 30 of Exp. 2. The lower asymptote levels for both Exp. 1 and 2 indicated highly depleted in ¹⁵N NH₃-N at the beginning of the manure storage process. The length of Exp. 1, however, was apparently too short to clearly observe the point of equilibrium visible in Exp. 2 (δ¹⁵N upper asymptote: 19.8‰). The inflexion point (i.e., the point of maximum rate of δ¹⁵N increase) was around day 13 and day 16 for Exp. 1 and Exp. 2, respectively. Delta ¹⁵N of manure reached a plateau beyond day 6 (Figure 2, Panel A), which coincided with the decline in NH₃-N emission rates. Although changes in δ¹⁵N of manure parallel NH₃-N losses, the dynamics of δ¹⁵N of NH₃-N will likely make application of the N isotope approach for estimating manure NH₃-N emissions difficult in practical farm conditions, which was our original goal [4]. Nevertheless, the relationship between manure δ¹⁵N and NH₃-N volatilization losses appears to be linear in the initial stages of manure storage, when NH₃-N losses are most intensive, and deserves further investigation.

The discrepancy between net manure N losses and NH₃-N recovered in the acid trap was further investigated in Exp. 3. By labeling the urea N pool with ¹⁵N, the main source of emitted NH₃-N could be traced. Results of this experiment are shown in Figure 3. Nitrogen-15 enrichment of both manure- and NH₃-N pools rapidly declined within 5 days of the incubation (Figure 3), representing the most intensive phase of NH₃-N losses. For the manure-N pool, the ¹⁵N decay clearly reflected loss of highly enriched in ¹⁵N volatile N. The decline in ¹⁵N-enrichment of the NH₃-N pool followed the ¹⁵N decay of the source manure N pool and reflected the rapid decline in urea-N concentration observed in these experiments (Figure 1B). The decline in ¹⁵N-enrichment, however, was much more rapid for manure- compared with NH₃-N (rate constants of 1.473 and 0.385 APE/day, respectively). This would clearly represent ¹⁵N loss other than NH₃-N, which is in agreement with the suggested large N₂ loss in the initial hours of manure storage (see earlier discussion) and is supported by the studies of Harper et al. [3] with swine manure. Although isotope fractionation and discrimination against the heavier N isotope, as reported for Exp. 1 and 2, were undoubtedly taking place in Exp. 3, the δ¹⁵N values of the manure-N pool in this experiment (δ¹⁵N 4306 ± 77.7‰) was so much greater than the natural abundance of ¹⁵N in manure (δ¹⁵N ‒0.51± 0.42‰; Exp. 1 and 2) that these processes could not have had a measurable impact on the ¹⁵N-enrichment data from Exp. 3. The absolute losses of ¹⁵N during the 20 day simulated manure storage were on average 68.2 ± 2.45 mg. This represented approximately 70% of the 98 mg urea-¹⁵N introduced into each incubation vessel at day 0 (200 mg of 98 atom % ¹⁵N-urea). The amount of ¹⁵N recovered as NH₃-N was 51.2 ± 1.85 mg, or 75.2 ± 4.04% of the urea-¹⁵N lost in 20 days as NH₃-N. The difference of urea-¹⁵N lost and trapped as NH₃-N (about 25%) is supportive of the hypothesis that gaseous N losses other than NH₃ may be responsible for part of manure N losses during storage. This process is likely taking place exclusively in the initial days of manure storage. Averaged manure-¹⁵N and NH₃-¹⁵N losses data were fitted to a non-linear model (double exponential decay; data not shown) and the predicted value were used to calculate urea-¹⁵N recovery as NH₃-N during the initial 3 days of manure storage. As expected, recovery of daily urea-¹⁵N lost from manure as NH₃-N was the lowest at day 1 (30.4%) and day 2 (45.4%) of the incubation; recovery was complete (102.7%) by day 3. This trend supports the concept that volatile nitrogenous compounds other than NH₃ (likely N₂) could account for as much as 50 to 70% of the N losses during the initial 48 h of cattle manure storage. Harper et al. [3] reported 2 to 8 times greater N₂ than NH₃-N emissions from swine lagoons in Georgia and North Carolina. Dinitrogen gas emission would be still dependent on manure composition, specifically urinary urea excretion by the animal; lagoon ammonium concentration was the primary factor determining N2 emissions in the Harper et al. [3] study. These authors concluded that swine lagoons emit much less NH₃-N than previously estimated. Based on results from the current study, similar conclusion may be drawn for cattle manure. Our laboratory data, indicating about 25% manure N losses unaccounted as NH₃-N (in 20 days), need to be confirmed in field experiments, where a multitude of environmental factors affect N volatilization losses from dairy and beef cattle operations [2].

One of the objectives of this study was to quantify the contribution of urinary- vs. fecal-N to NH₃-N emitted from cattle manure. In Exp. 4, 2 types of manure were produced: FLM, ¹⁵N-labeled feces and unlabeled urine and ULM, unlabeled feces and ¹⁵N-labeled urine. Labeled and unlabeled feces or urine used to prepare FLM and ULM had similar N concentrations: on average 0.48 ± 0.01 and 1.01 ± 0.02%, respectively. As a result, both types of manure had similar (P = 0.14 to 0.44) DM (data not shown), N, and urea-N concentrations (Table 2). Consequently, cumulative or daily NH₃-N emissions were also not different (P = 0.59) between the 2 treatments. The goal of labeling feces or urine with ¹⁵N was successfully achieved. Delta ¹⁵N of feces in FLM manure was approximately 17-times higher (P = 0.022) than that of feces in ULM manure. Similarly, δ¹⁵N of urine in ULM manure was drastically higher (P < 0.001) than that of urine in FLM, which resulted in δ¹⁵N of ULM being higher (P < 0.001) than δ¹⁵N of FLM. Delta ¹⁵N of unlabeled feces and urine for both types of manure was within the range of natural δ¹⁵N reported for dairy cows [4]. Nitrogen-15 enrichment of NH₃-N followed a sigmoid trend for FLM and an exponential decay trend for ULM (adjusted R² = 0.86 and 0.92, P = 0.005 and < 0.001, respectively; Figure 4A). The estimated proportion of NH₃-N originating from fecal N (FLM) was negligible in the first 48 h of manure storage, represented 0.04 ± 0.006 by day 5, and then gradually increased to 0.11 ± 0.019 of the emitted NH₃-N by day 10 (logistic regression model; adjusted R² = 0.91, P < 0.001) (Figure 4B). The proportion of NH3-N originating from urinary N (ULM) represented 0.94 ± 0.027 at 24 h, 0.97 ± 0.002 at 48 h, 0.91 ± 0.004 at 72 h, and gradually decreased to 0.87 ± 0.005 by day 10 (exponential decay model; adjusted R² = 0.92, P < 0.001). This experiment clearly identified urinary N as the principal source of NH₃-N volatilized from cattle manure during the initial 10 days of storage, accounting for an average of 90% of the emitted NH₃-N. The contribution of fecal N was relatively low, but gradually increased to about 10% by day 10. Using a similar approach, Thomsen [7] estimated that urinary N accounted for 79% of the total N losses from sheep manure after 7 days of composting, decreasing to 64% at the end of the 86-day storage period. In manure stored anaerobically, urinary N accounted for 94% of the total N losses after 28 days and for 68% at 86 days [7].

These series of laboratory experiments confirmed that NH3 volatilization losses from manure are most intensive during the first 2 to 3 days of manure storage and this coincides with a rapid loss (hydrolysis) of urinary urea. The relationship between δ¹⁵N of manure and NH₃ emission appears to be linear during the initial stages of manure storage (when most of the NH3 losses occur) and should be further investigated. The main source of NH₃-N volatilized from cattle manure during the initial 10 days of storage is urinary-N, representing on average 90% of the emitted NH₃-N. The contribution of fecal N was relatively low, but increased to about 10% by day 10. There appears to be substantial emissions of nitrogenous gases other than NH₃, most likely dinitrogen gas, which may account for up to 25% of N losses during the first 20 day of manure storage. This finding, which has to be confirmed in laboratory and field experiments, may be indicative of overestimation of NH3 emissions from cattle operations by the current emissions factors.